Traffic Density Estimation Using Vehicle ADAS Data

Abstract: Traffic density is the fundamental and crucial data in traffic analysis process. Therefore, traffic density estimation has been an important research topicin traffic engineering field. As a means of collecting the fundamental data, loop detectors have been used over a large area for a long time and haveshown satisfactory performance results. However, conventional detectors have difficulties because of the high cost of installation and maintenance.Therefore, a number of methods for replacing loop detectors have been studied, however most of them still request investment of additionalinfrastructure to collect traffic density data. This is why traffic density is important data but difficult to collect and utilize.

In recent years, new opportunities to collect traffic data are arising with the increasing number of vehicles equipped with safety related devices such asAdvanced Driver Assistance Systems (ADAS). Moreover, in case of commercial vehicles, it is mandatory to install a video recording device for thepurpose of safety management in Korea, which means that it is easier to secure probe vehicles for data collection than the past.

Based on this idea, this study developed traffic density estimation method for multilane highway using vehicle ADAS data. As the first step, this studycollected the distance to the target vehicles from the ADAS equipped vehicle. To get more data sample of distance, this study used the probe vehicleequipped with front and rear ADAS to detect both preceding and trailing vehicle. With these distance data, this study calculated traffic density at thesecond step. In this step, this study adjusted space and time resolution in order to get more accurate density data. Also, this study reflectedcharacteristics of multilane highways that have both interrupted and uninterrupted traffic flow in the process of traffic density estimation. Finally, thisstudy evaluated the proposed method based on Mean Absolute Deviation (MAD), Mean Square Error (MSE), and Mean Absolute Percentage Error(MAPE) as evaluation measures.